KMeans Clustering Analysis

Authors

Affiliation

Ling Lu

Boston University

Luoyan Zhang

Boston University

Yinuo Wang

Boston University

Published

April 21, 2025

Load the Dataset

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.cluster import KMeans
from sklearn.metrics import confusion_matrix

pd.set_option('display.max_columns', None)
df = pd.read_csv("lightcast_job_postings.csv")

df.info()
df.head()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 72498 entries, 0 to 72497
Columns: 118 entries, id to naics_2022_6_name
dtypes: float64(38), object(80)
memory usage: 65.3+ MB

	id	last_updated_date	last_updated_timestamp	duplicates	posted	expired	duration	title_raw	body	modeled_expired	modeled_duration	company	company_name	company_raw	company_is_staffing	education_levels	education_levels_name	min_edu_levels	min_edu_levels_name	max_edu_levels	max_edu_levels_name	employment_type	employment_type_name	min_years_experience	max_years_experience	is_internship	salary	remote_type	remote_type_name	original_pay_period	salary_to	salary_from	location	city	city_name	county	county_name	msa	msa_name	state	state_name	county_outgoing	county_name_outgoing	county_incoming	county_name_incoming	msa_outgoing	msa_name_outgoing	msa_incoming	msa_name_incoming	naics2	naics2_name	naics3	naics3_name	naics4	naics4_name	naics5	naics5_name	naics6	naics6_name	title	title_name	title_clean	certifications	certifications_name	onet	onet_name	onet_2019	onet_2019_name	cip6	cip6_name	cip4	cip4_name	cip2	cip2_name	soc_2021_2	soc_2021_2_name	soc_2021_3	soc_2021_3_name	soc_2021_4	soc_2021_4_name	soc_2021_5	soc_2021_5_name	lot_career_area	lot_career_area_name	lot_occupation	lot_occupation_name	lot_specialized_occupation	lot_specialized_occupation_name	lot_occupation_group	lot_occupation_group_name	lot_v6_specialized_occupation	lot_v6_specialized_occupation_name	lot_v6_occupation	lot_v6_occupation_name	lot_v6_occupation_group	lot_v6_occupation_group_name	lot_v6_career_area	lot_v6_career_area_name	soc_2	soc_2_name	soc_3	soc_3_name	soc_4	soc_4_name	soc_5	soc_5_name	lightcast_sectors	lightcast_sectors_name	naics_2022_2	naics_2022_2_name	naics_2022_3	naics_2022_3_name	naics_2022_4	naics_2022_4_name	naics_2022_5	naics_2022_5_name	naics_2022_6	naics_2022_6_name
0	1f57d95acf4dc67ed2819eb12f049f6a5c11782c	9/6/2024	32:57.4	0.0	6/2/2024	6/8/2024	6.0	Enterprise Analyst (II-III)	31-May-2024\n\nEnterprise Analyst (II-III)\n\n...	False	6.0	894731.0	Murphy USA	Murphy USA	False	[\n 2\n]	[\n "Bachelor's degree"\n]	2.0	Bachelor's degree	NaN	NaN	1.0	Full-time (> 32 hours)	2.0	2.0	False	NaN	0.0	[None]	NaN	NaN	NaN	{\n "lat": 33.20763,\n "lon": -92.6662674\n}	RWwgRG9yYWRvLCBBUg==	El Dorado, AR	5139.0	Union, AR	20980.0	El Dorado, AR	5.0	Arkansas	5139.0	Union, AR	5139.0	Union, AR	20980.0	El Dorado, AR	20980.0	El Dorado, AR	44.0	Retail Trade	441.0	Motor Vehicle and Parts Dealers	4413.0	Automotive Parts, Accessories, and Tire Retailers	44133.0	Automotive Parts and Accessories Retailers	441330.0	Automotive Parts and Accessories Retailers	ET29C073C03D1F86B4	Enterprise Analysts	enterprise analyst ii iii	[]	[]	15-2051.01	Business Intelligence Analysts	15-2051.01	Business Intelligence Analysts	[\n "45.0601",\n "27.0101"\n]	[\n "Economics, General",\n "Mathematics, Ge...	[\n "45.06",\n "27.01"\n]	[\n "Economics",\n "Mathematics"\n]	[\n "45",\n "27"\n]	[\n "Social Sciences",\n "Mathematics and St...	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	23.0	Information Technology and Computer Science	231010.0	Business Intelligence Analyst	23101011.0	General ERP Analyst / Consultant	2310.0	Business Intelligence	23101011.0	General ERP Analyst / Consultant	231010.0	Business Intelligence Analyst	2310.0	Business Intelligence	23.0	Information Technology and Computer Science	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	[\n 7\n]	[\n "Artificial Intelligence"\n]	44.0	Retail Trade	441.0	Motor Vehicle and Parts Dealers	4413.0	Automotive Parts, Accessories, and Tire Retailers	44133.0	Automotive Parts and Accessories Retailers	441330.0	Automotive Parts and Accessories Retailers
1	0cb072af26757b6c4ea9464472a50a443af681ac	8/2/2024	08:58.8	0.0	6/2/2024	8/1/2024	NaN	Oracle Consultant - Reports (3592)	Oracle Consultant - Reports (3592)\n\nat SMX i...	False	NaN	133098.0	Smx Corporation Limited	SMX	True	[\n 99\n]	[\n "No Education Listed"\n]	99.0	No Education Listed	NaN	NaN	1.0	Full-time (> 32 hours)	3.0	3.0	False	NaN	1.0	Remote	NaN	NaN	NaN	{\n "lat": 44.3106241,\n "lon": -69.7794897\n}	QXVndXN0YSwgTUU=	Augusta, ME	23011.0	Kennebec, ME	12300.0	Augusta-Waterville, ME	23.0	Maine	23011.0	Kennebec, ME	23011.0	Kennebec, ME	12300.0	Augusta-Waterville, ME	12300.0	Augusta-Waterville, ME	56.0	Administrative and Support and Waste Managemen...	561.0	Administrative and Support Services	5613.0	Employment Services	56132.0	Temporary Help Services	561320.0	Temporary Help Services	ET21DDA63780A7DC09	Oracle Consultants	oracle consultant reports	[]	[]	15-2051.01	Business Intelligence Analysts	15-2051.01	Business Intelligence Analysts	[]	[]	[]	[]	[]	[]	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	23.0	Information Technology and Computer Science	231010.0	Business Intelligence Analyst	23101012.0	Oracle Consultant / Analyst	2310.0	Business Intelligence	23101012.0	Oracle Consultant / Analyst	231010.0	Business Intelligence Analyst	2310.0	Business Intelligence	23.0	Information Technology and Computer Science	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	NaN	NaN	56.0	Administrative and Support and Waste Managemen...	561.0	Administrative and Support Services	5613.0	Employment Services	56132.0	Temporary Help Services	561320.0	Temporary Help Services
2	85318b12b3331fa490d32ad014379df01855c557	9/6/2024	32:57.4	1.0	6/2/2024	7/7/2024	35.0	Data Analyst	Taking care of people is at the heart of every...	False	8.0	39063746.0	Sedgwick	Sedgwick	False	[\n 2\n]	[\n "Bachelor's degree"\n]	2.0	Bachelor's degree	NaN	NaN	1.0	Full-time (> 32 hours)	5.0	NaN	False	NaN	0.0	[None]	NaN	NaN	NaN	{\n "lat": 32.7766642,\n "lon": -96.7969879\n}	RGFsbGFzLCBUWA==	Dallas, TX	48113.0	Dallas, TX	19100.0	Dallas-Fort Worth-Arlington, TX	48.0	Texas	48113.0	Dallas, TX	48113.0	Dallas, TX	19100.0	Dallas-Fort Worth-Arlington, TX	19100.0	Dallas-Fort Worth-Arlington, TX	52.0	Finance and Insurance	524.0	Insurance Carriers and Related Activities	5242.0	Agencies, Brokerages, and Other Insurance Rela...	52429.0	Other Insurance Related Activities	524291.0	Claims Adjusting	ET3037E0C947A02404	Data Analysts	data analyst	[\n "KS683TN76T77DQDVBZ1B"\n]	[\n "Security Clearance"\n]	15-2051.01	Business Intelligence Analysts	15-2051.01	Business Intelligence Analysts	[]	[]	[]	[]	[]	[]	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	23.0	Information Technology and Computer Science	231113.0	Data / Data Mining Analyst	23111310.0	Data Analyst	2311.0	Data Analysis and Mathematics	23111310.0	Data Analyst	231113.0	Data / Data Mining Analyst	2311.0	Data Analysis and Mathematics	23.0	Information Technology and Computer Science	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	NaN	NaN	52.0	Finance and Insurance	524.0	Insurance Carriers and Related Activities	5242.0	Agencies, Brokerages, and Other Insurance Rela...	52429.0	Other Insurance Related Activities	524291.0	Claims Adjusting
3	1b5c3941e54a1889ef4f8ae55b401a550708a310	9/6/2024	32:57.4	1.0	6/2/2024	7/20/2024	48.0	Sr. Lead Data Mgmt. Analyst - SAS Product Owner	About this role:\n\nWells Fargo is looking for...	False	10.0	37615159.0	Wells Fargo	Wells Fargo	False	[\n 99\n]	[\n "No Education Listed"\n]	99.0	No Education Listed	NaN	NaN	1.0	Full-time (> 32 hours)	3.0	NaN	False	NaN	0.0	[None]	NaN	NaN	NaN	{\n "lat": 33.4483771,\n "lon": -112.0740373\n}	UGhvZW5peCwgQVo=	Phoenix, AZ	4013.0	Maricopa, AZ	38060.0	Phoenix-Mesa-Chandler, AZ	4.0	Arizona	4013.0	Maricopa, AZ	4013.0	Maricopa, AZ	38060.0	Phoenix-Mesa-Chandler, AZ	38060.0	Phoenix-Mesa-Chandler, AZ	52.0	Finance and Insurance	522.0	Credit Intermediation and Related Activities	5221.0	Depository Credit Intermediation	52211.0	Commercial Banking	522110.0	Commercial Banking	ET2114E0404BA30075	Management Analysts	sr lead data mgmt analyst sas product owner	[]	[]	15-2051.01	Business Intelligence Analysts	15-2051.01	Business Intelligence Analysts	[]	[]	[]	[]	[]	[]	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	23.0	Information Technology and Computer Science	231113.0	Data / Data Mining Analyst	23111310.0	Data Analyst	2311.0	Data Analysis and Mathematics	23111310.0	Data Analyst	231113.0	Data / Data Mining Analyst	2311.0	Data Analysis and Mathematics	23.0	Information Technology and Computer Science	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	[\n 6\n]	[\n "Data Privacy/Protection"\n]	52.0	Finance and Insurance	522.0	Credit Intermediation and Related Activities	5221.0	Depository Credit Intermediation	52211.0	Commercial Banking	522110.0	Commercial Banking
4	cb5ca25f02bdf25c13edfede7931508bfd9e858f	6/19/2024	00:00.0	0.0	6/2/2024	6/17/2024	15.0	Comisiones de $1000 - $3000 por semana... Comi...	Comisiones de $1000 - $3000 por semana... Comi...	False	15.0	0.0	Unclassified	LH/GM	False	[\n 99\n]	[\n "No Education Listed"\n]	99.0	No Education Listed	NaN	NaN	3.0	Part-time / full-time	NaN	NaN	False	92500.0	0.0	[None]	year	150000.0	35000.0	{\n "lat": 37.6392595,\n "lon": -120.9970014\n}	TW9kZXN0bywgQ0E=	Modesto, CA	6099.0	Stanislaus, CA	33700.0	Modesto, CA	6.0	California	6099.0	Stanislaus, CA	6099.0	Stanislaus, CA	33700.0	Modesto, CA	33700.0	Modesto, CA	99.0	Unclassified Industry	999.0	Unclassified Industry	9999.0	Unclassified Industry	99999.0	Unclassified Industry	999999.0	Unclassified Industry	ET0000000000000000	Unclassified	comisiones de por semana comiensa rapido	[]	[]	15-2051.01	Business Intelligence Analysts	15-2051.01	Business Intelligence Analysts	[]	[]	[]	[]	[]	[]	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	23.0	Information Technology and Computer Science	231010.0	Business Intelligence Analyst	23101012.0	Oracle Consultant / Analyst	2310.0	Business Intelligence	23101012.0	Oracle Consultant / Analyst	231010.0	Business Intelligence Analyst	2310.0	Business Intelligence	23.0	Information Technology and Computer Science	15-0000	Computer and Mathematical Occupations	15-2000	Mathematical Science Occupations	15-2050	Data Scientists	15-2051	Data Scientists	NaN	NaN	99.0	Unclassified Industry	999.0	Unclassified Industry	9999.0	Unclassified Industry	99999.0	Unclassified Industry	999999.0	Unclassified Industry

Data Cleaning

selected_features = [
    'remote_type_name', 
    'state_name', 
    'city_name', 
    'lot_occupation_name',
    'min_edu_levels_name', 
    'salary'
]

X = df[selected_features]

X = X.dropna()

salary = X['salary']
X_others = X.drop(columns=['salary'])

X_encoded = pd.get_dummies(X_others)

scaler = StandardScaler()
salary_scaled = scaler.fit_transform(salary.values.reshape(-1, 1))

X_final = np.concatenate([X_encoded.values, salary_scaled], axis=1)

X_encoded.index = X.index

Train/Test Split

X_train, X_test, train_idx, test_idx = train_test_split(
    X_final, X_encoded.index, test_size=0.2, random_state=688
)

Finding K using Elbow Method

inertia = []
K_range = range(1, 11)

for k in K_range:
    kmeans = KMeans(n_clusters=k, random_state=42)
    kmeans.fit(X_train)
    inertia.append(kmeans.inertia_)

plt.plot(K_range, inertia, marker='o')
plt.xlabel('Number of clusters (K)')
plt.ylabel('Inertia')
plt.title('Elbow Method For Optimal K')
plt.show()

Insights_ Elbow Method (K-Means Clustering)

In the plot above, we applied the Elbow Method to determine the optimal number of clusters (K) for our K-Means clustering model. The Y-axis represents the inertia, which is the sum of squared distances from each point to its assigned cluster center. The X-axis shows different values of K, ranging from 1 to 10.

We can observe that inertia decreases as K increases, which is expected — more clusters naturally lead to tighter groupings and lower within-cluster variance. However, the key insight from the elbow method is to look for the point where the rate of inertia reduction starts to slow down, forming an “elbow” shape. In our case, this elbow appears around K = 3 or 4, where the curve starts to flatten noticeably. Choosing a K beyond this point leads to only marginal improvements in inertia, suggesting diminishing returns.

Based on this analysis, we decided to use 4 clusters, which would likely balance interpretability and model effectiveness, helping us segment our data meaningfully without overfitting. This insight guides how we might group jobs, regions, or skills in our dataset for further analysis.

optimal_k = 4

Fit and Evaluate the Model

kmeans = KMeans(n_clusters=optimal_k, random_state=688)
train_clusters = kmeans.fit_predict(X_train)

true_labels_train = df.loc[train_idx, 'lot_occupation_name']

le = LabelEncoder()
true_labels_encoded = le.fit_transform(true_labels_train)

conf_mat = confusion_matrix(true_labels_encoded, train_clusters)

plt.figure(figsize=(12, 10))
sns.heatmap(conf_mat, annot=True, fmt='d', cmap='Blues')
plt.xlabel('Predicted Cluster')
plt.ylabel('True ONET Label')
plt.title('Confusion Matrix: KMeans vs ONET Classification')
plt.show()

Insights_Confusion Matrix: KMeans vs. ONET Classification:

This confusion matrix compares the predicted clusters from our KMeans model with the true job role classifications based on ONET labels. Each row represents the actual ONET category, while each column shows how many jobs from that category were assigned to a given KMeans cluster. From the visualization, we can see that certain ONET labels aligned strongly with specific clusters—for instance, ONET label 1 is heavily concentrated in Cluster 0, with over 7,500 records, suggesting strong cohesion in that group. Similarly, ONET label 4 dominates Clusters 2 and 3, showing that KMeans successfully identified distinct subgroups within that label.

However, other labels such as label 3 appear more scattered, indicating overlap between job types or fuzzier cluster boundaries. Interestingly, Clusters 4 and 5 were unused, suggesting that although we specified six clusters, only four were meaningfully populated. This may imply either redundancy in our chosen K value or limited variation in the feature space.

Overall, this matrix provides a valuable check on how well our unsupervised clustering aligns with known job categories, revealing both strengths in classification and areas for improvement in feature design or model tuning.

Choropleth Map of Average Salary by U.S. State

import plotly.express as px
import us


salary_by_state = X.groupby("state_name")["salary"].mean().sort_values(ascending=False).reset_index()
salary_by_state['state_name'] = salary_by_state['state_name'].apply(
    lambda x: us.states.lookup(x).abbr if us.states.lookup(x) else x
)

fig = px.choropleth(
    salary_by_state,
    locations='state_name',
    locationmode='USA-states',
    color='salary',
    color_continuous_scale='Blues',
    scope='usa',
    labels={'avg_salary': 'Avg Salary'},
    title='Average Salary by U.S. State',
       width=1000,  
     height=600  
)

fig.update_layout(coloraxis_colorbar=dict(
    title="Avg Salary",
    tickprefix="$",
    ticks="outside"
))

fig.show()

Insights_Choropleth Map of Average Salary by U.S. State:

This choropleth map visualizes the average salary by state across the United States, with darker shades of blue indicating higher average compensation levels. From the visualization, we can observe that states such as California, New York, and Washington D.C. offer the highest average salaries, aligning with their large urban labor markets, dense concentrations of tech companies, and elevated costs of living.

In contrast, many southern and central states like Mississippi, New Mexico, and West Virginia report lower average salaries, reflecting a mix of lower living costs and possibly fewer high-paying industry hubs. As a student analyzing this data, this visualization emphasizes how geographic location significantly influences earning potential. It also highlights opportunities for remote work seekers to access high-paying markets without relocation, provided employers support flexible arrangements.

This map supports our broader analysis by revealing where economic opportunities are most concentrated, which can inform personal job strategies or workforce development planning.

Salary Variation Across KMeans Clusters

X_train_clustered = pd.DataFrame(X_train, index=train_idx)
X_train_clustered['Cluster'] = train_clusters
X_train_clustered['state_name'] = df.loc[train_idx, 'state_name']
X_train_clustered['remote_type_name'] = df.loc[train_idx, 'remote_type_name']
X_train_clustered['salary'] = df.loc[train_idx, 'salary']

X_train_clustered.head()

	1	2	3	12	27	31	36	46	404	1024	1145	1752	2037	2368	2370	2371	2373	2374	2377	2379	Cluster	state_name	remote_type_name	salary
59430	0.0	1.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	1.0	0.0	2.029401	1	Mississippi	Remote	209547.0
60793	1.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	1.0	0.0	1.0	0.0	0.0	0.0	1.0	0.0	-1.306219	2	Texas	Not Remote	59000.0
60680	0.0	0.0	1.0	0.0	0.0	0.0	1.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	1.0	0.0	0.290169	0	North Carolina	[None]	131050.0
18925	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	1.0	0.0	0.0	0.0	1.0	0.612549	1	Florida	[None]	145600.0
7237	0.0	1.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	1.0	0.0	0.0	1.0	0.0	0.0	-1.128943	3	Nevada	Remote	67001.0

import seaborn as sns
import matplotlib.pyplot as plt

plt.figure(figsize=(10, 6))
sns.boxplot(x='Cluster', y='salary', data=X_train_clustered)
plt.title('Salary Distribution by Cluster')
plt.xlabel('Cluster')
plt.ylabel('Salary')
plt.show()

Insights_Salary Variation Across KMeans Clusters

In this section, we analyzed how salaries vary across the job clusters produced by our KMeans model. The boxplot visualizes the distribution of salaries within each cluster (Cluster 0 to Cluster 3), allowing us to observe central tendencies, spread, and outliers. We can clearly see that Cluster 1 has the highest median salary, with a large number of high-earning outliers, suggesting it may represent higher-skilled or more specialized job roles. Clusters 0, 2, and 3 have comparatively lower median salaries, with Cluster 3 showing the lowest central value and the narrowest salary range. This suggests that Cluster 3 likely contains more entry-level or uniform-pay positions.

From our perspective, this analysis helps us understand the economic implications of unsupervised clustering. Even though we didn’t use salary as a feature to form the clusters, we see meaningful differences in pay across them. These differences validates that the clusters may capture latent job characteristics (e.g., seniority, technical skill, or industry) that correlate with compensation. It also suggests that further profiling of these clusters using features like remote_type_name or state_name could provide deeper insight into which job segments drive higher salaries.

State-Level Job Distribution Across Clusters

top_states = X_train_clustered['state_name'].value_counts().nlargest(10).index
subset = X_train_clustered[X_train_clustered['state_name'].isin(top_states)]

plt.figure(figsize=(12, 6))
sns.countplot(x='state_name', hue='Cluster', data=subset)
plt.title('Top 10 States Job Distribution by Cluster')
plt.xlabel('State')
plt.ylabel('Job Count')
plt.xticks(rotation=45)
plt.legend(title='Cluster')
plt.show()

The Kernel crashed while executing code in the current cell or a previous cell. 

Please review the code in the cell(s) to identify a possible cause of the failure. 

Click <a href='https://aka.ms/vscodeJupyterKernelCrash'>here</a> for more info. 

View Jupyter <a href='command:jupyter.viewOutput'>log</a> for further details.

Insights_State-Level Job Distribution Across Clusters

This bar chart displays the job distribution across the top 10 U.S. states by KMeans cluster, helping us understand how different job types (as identified by clustering) are spread geographically. Each bar represents the number of job postings in a particular state, broken down by cluster label (0 through 3).

We observe that Texas dominates in job count across all clusters, especially Cluster 0, suggesting it is a highly diversified job market that spans multiple segments. North Carolina and Florida also show noticeable volumes, with Florida particularly rich in Cluster 2 jobs — which, based on earlier salary distribution analysis, may include mid- to lower-wage roles. California, while often associated with high-paying jobs, shows a more balanced distribution across Clusters 1, 2, and 3, implying a broader range of roles.

As we exploring geographic job trends, this visualization gives us insight into how job market segmentation varies by region. States like Texas and Florida seem to attract or offer a wide range of clustered job types, while states like New York or Washington have a more balanced but smaller presence across clusters. This supports the idea that location influences not just salary, but also the types of roles available, and could be a key factor for workforce planning or job-seeking strategies.

Job Seeker Recommendation:

Based on our salary distribution boxplot by cluster (see: Salary Distribution by Cluster), job seekers should focus on roles that fall into Cluster 1, which has the highest median salary and the widest spread of high-paying outliers. This suggests that Cluster 1 contains more senior or in-demand job types. To position themselves for these opportunities, candidates can review the common job titles and education levels within this cluster (as shown in the cluster composition plots by education and occupation) and tailor their resumes or training plans accordingly.

Additionally, the state-level cluster distribution bar chart (Top 10 States Job Distribution by Cluster) shows that states like Texas and Florida offer a wide range of job opportunities across all clusters. Job seekers may consider these locations for broader access to diverse roles, especially if they’re open to relocating or seeking hybrid positions.